Public-Safe Explanations

Generate explanations that avoid private internals and accusation-heavy language.

Public-safe explanations are written for shareable contexts. They avoid restricted traces, internal diagnostic wording, and certain-judgment language.

The current evidence is internal/local report hygiene, not legal, compliance, or external review.

Current evidence combines internal/local tests and controlled hosted API smoke. External validation and production hardening are separate future milestones.